“…In addition to issues common to PET imaging itself, such as noise or partial volume effects [ 47 , 48 , 49 ], radiomics must deal with standardization issues related to the differences in acquisition and reconstruction parameters, post-processing techniques, tumor segmentation methods or even texture calculations [ 46 , 50 , 51 , 52 , 53 ]. The complex formulation of radiomics makes it challenging to explain pervasive findings such as the correlations displayed between different texture indices [ 54 , 55 , 56 , 57 , 58 , 59 , 60 ], as well as strong correlations between textural indices and tumor volume [ 61 , 62 , 63 ], which compromise the added value of these parameters in comparison with SUV or MTV. Furthermore, the variability in definitions and nomenclature of heterogeneity metrics themselves complicate any evaluation and comparison of published results [ 64 ].…”